Date: Thu, 05 Apr 2001 15:16:05 -0400 From: "Louis A. Mamakos" <louie@TransSys.COM> Cc: Wes Peters <wes@softweyr.com>, Brett Glass <brett@lariat.org>, freebsd-net@FreeBSD.ORG Subject: Re: Transition from modem PPP to PPPoE Message-ID: <200104051916.f35JG5n54176@whizzo.transsys.com> In-Reply-To: Your message of "Thu, 05 Apr 2001 13:38:54 EDT." <200104051738.f35Hcsn53390@whizzo.transsys.com> References: <4.3.2.7.2.20010330201802.00dc8f00@localhost> <4.3.2.7.2.20010401141552.0452a6c0@localhost> <3ACBF0B6.52B99863@softweyr.com> <200104051738.f35Hcsn53390@whizzo.transsys.com>
next in thread | previous in thread | raw e-mail | index | archive | help
> I've never thought that the 4 bytes of overhead per PPPoE frame was > terribly inefficient, compared to, say, IP-in-IP with another 20 byte > IP header. But I'm certainly not arguing that a choice of technology > be made on simply the number of bytes on the wire; there are other > things to consider as well. Ooops, must have been smoking some of Jordan's crack. That's more like 14 bytes rather than 4. Still, we're in the same ballpark at other schemes of tunneling over ethernet. I think that the code path-length might be a bit longer for PPPoE, but that's a wild-ass guess. I have a suspicion that there's slightly more overhead paid for the netgraph-based implementation as compared to a optimally coded IP-in-IP tunnel (using gif?). Of course, the netgraph implementation is a huge win over running the packets up into user mode and doing a context switch. louie To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200104051916.f35JG5n54176>